According to the latest research report disclosed by GF Securities, a sell-side research institution, Apple is planning to introduce Intel's most advanced foundry process nodes, including 18A-P and 14A, in future generations of self-developed chips for processor production in different product lines. Reports indicate that Apple will use Intel's 18A-P process for the M7 series of system-on-chip (SoC), which will provide computing power for notebook products such as MacBook Air and entry-level MacBook Pro. At the same time, Intel is increasing its investment in R&D and mass production of the 14A node, and Apple plans to use this process in the future to manufacture the A21 chip for the new generation of iPhones.
Citing previously public information, the article stated that compared to the standard 18A process, the 18A-P node can bring about a 9% performance improvement at the same power consumption, or reduce power consumption by about 18% at the same performance level. This balance of performance and energy efficiency is considered very suitable for use in notebook SoCs for thin and light notebooks and mainstream productivity notebooks, and is expected to bring higher operating frequencies and lower energy consumption to the new generation of M7. As Apple gradually migrates from the TSMC 3nm process node used in the current M5 chip, the industry expects that with the support of the new process, the new MacBook series will usher in a significant upgrade in terms of performance and battery life. Relevant changes are expected to be gradually reflected in end products around 2027.
In the field of smartphones, Apple is accused of planning to use Intel's 14A process for future A21 SoC. The report believes that the 14A node is expected to achieve an "intergenerational leap" in terms of transistor density, frequency potential and power consumption performance, which is in line with Apple's long-term goal of pursuing higher performance and longer battery life on mobile devices. Apple's current timeline is to officially launch iPhones equipped with the 14A process A21 chip by 2028. Since this process still takes about two years to prepare, Apple is likely to wait for the final PDK (process design kit) of the 14A process to be finalized before starting trial production and tape-out verification of the chip.
It is worth noting that it is unclear whether Apple will adopt a "dual-source foundry" strategy, that is, the high-end version A21 Pro will continue to be produced by TSMC, while the regular version of the A21 will be handed over to Intel. Regardless of the specific plan, it is generally believed that Apple is intending to gradually diversify its supply chain in the high-end chip field and no longer rely entirely on a single wafer factory. In terms of the layout of key links such as advanced packaging, Intel has continued to increase its investment in recent years, allowing it to compete with TSMC in some areas. Apple's move is also seen as a positive response to this trend.
From the perspective of manufacturing and packaging processes, reports indicate that in order to meet the performance and energy efficiency goals of the M7 SoC, Apple’s solution is likely to require a combination of advanced packaging technologies. This includes various forms of Intel's Foveros packaging family, such as Foveros-S, Foveros-R, Foveros-B or Foveros Direct, coupled with technologies such as EMIB (Embedded Multi-chip Interconnect Bridge). The Foveros solution can provide more flexible multi-chip packaging through interposers and redistribution layers (RDL), while supporting true 3D stacking through copper-to-copper hybrid bonding to meet application scenarios with extremely high inter-die bandwidth or extreme energy efficiency.
In terms of EMIB, Intel not only provides conventional small-size silicon intermediary bridges, but also extends a variety of variants, such as EMIB-M with integrated metal-insulated-metal (MIM) capacitors, and EMIB-T with through-silicon vias (TSV). These technology combinations can help chips achieve more complex interconnect structures and higher signal integrity in small packages, providing more implementation paths for Apple's potential multi-chip SoC designs. Industry analysts believe that if the cooperation between the two parties is successfully implemented, in the next few years, the market is expected to see a batch of Apple's self-developed chip products with high performance, long battery life and complex packaging structures, which will also further heat up competition in the field of high-end processes and advanced packaging.